By default, AI is too polite. It nods along, agrees with your premises, and praises your prose. That is comfortable but useless for building real thinking. With the right prompts you can flip the dynamic: AI argues against you, finds holes in your reasoning, and forces you to defend every claim. That is when critical thinking actually grows.
Strong students, strong writers, and strong professionals all share one habit — they stress-test their own ideas before someone else does. Until recently, that required a smart friend, a tough teacher, or a generous colleague willing to disagree with you. AI gives you that interlocutor on demand, but only if you explicitly ask it to disagree. Otherwise its training pushes it toward agreement and encouragement.
This tutorial covers three modes: full structured debate, devil's-advocate stress-testing, and Socratic questioning. All three rely on the same trick: telling AI to stop being polite and start being rigorous.
Every argument has the same skeleton — a claim, the evidence backing it, the reasoning that connects evidence to claim, and the counter-arguments someone could raise. Weak arguments hide one or more of these. Strong arguments expose all four and address them.
A useful debate partner pokes at each of the four. They ask: what is your actual claim — in one sentence? What evidence do you have? How does that evidence support the claim — could the same evidence support a different conclusion? What is the strongest opposing view, and how do you answer it? AI can play that role for you on any topic, from your essay's thesis to your career choice, if you set it up to do so.
Weak prompt — gets a yes-man
I think social media has done more harm than
good to teenagers. What do you think?
AI will agree, restate your points more eloquently, and add a couple of supporting facts. You walk away feeling clever and learn nothing. Your argument is just as soft as it was before — and the day the examiner pushes back, you will not have a prepared answer.
Strong prompt — full debate setup
Act as a sharp, well-read debate partner. You
disagree with me on principle today.
Your job is to defend the position that
"social media has, on balance, helped teenagers
more than it has harmed them."
Rules:
1. Argue your position seriously, with concrete
examples and at least one piece of evidence
per claim. No vague generalities.
2. After your opening argument, I will respond.
Then attack my response — find the weakest
part of what I said and press on it.
3. Do not be polite. Be rigorous. If I make a
logical error (e.g. anecdotal evidence,
strawman, false dichotomy), name it.
4. After 4 exchanges, stop and write a
"verdict" paragraph: who actually had the
stronger argument, and why.
Begin with your opening argument.
Now you have a worthy opponent. Each exchange forces you to clarify your claim, sharpen your evidence, or admit you had not thought of something. That last "verdict" line is the gold — an honest external assessment of your reasoning.
List every logical fallacy or rhetorical trick in this argument. Name each one, quote the exact sentence, and explain why it is fallacious.
One of the fastest ways to learn the catalogue of common reasoning errors.
Based on this exchange, where was my reasoning strongest and where was it weakest? Be specific.Tip: The most useful exam preparation a humanities student can do is to debate the essay question with AI for thirty minutes before writing it. By the time you write the actual essay, every counter-argument is already in your head — and the examiner cannot surprise you.
Pick a position you genuinely hold on a current debate — climate policy, AI in classrooms, banning a book, anything you have an opinion on. Run the full debate prompt above. Note which of your arguments survived four exchanges, and which collapsed. Rewrite the survivors into a one-paragraph thesis.
Take an essay you wrote recently. Paste it with the prompt:
Act as a strict examiner. Identify every claim I make that is not backed by evidence, every logical jump, and every counter-argument I failed to address. List them as numbered weaknesses, most damaging first.
Use this to redraft.
Try the Socratic mode on a decision you are facing — choosing a degree, accepting a job offer, picking which competitive exam to prepare for. Prompt: "Do not give me advice. Ask me five questions, one at a time, that will help me think this decision through more clearly." The answers you struggle to articulate are exactly where you need to think harder.
Sign in to join the discussion and post comments.
Sign inPrompt Engineering for Specific AI Tools
Tool-by-tool mastery — deep dives into ChatGPT, Claude, Gemini, GitHub Copilot, Midjourney, Stable Diffusion, and more. Learn the exact prompting techniques each platform rewards.
Prompt Engineering for Developers
Use AI as your coding co-pilot. 18 tutorials on writing prompts to generate clean code, debug faster, write tests, build APIs, and ship better software.
Prompt Engineering for Content & Copywriting
Write blogs, ads, emails, and social media content ten times faster with AI. 13 practical tutorials on prompt engineering for content creators and copywriters.
Foundations of Prompt Engineering
The must-know basics of prompt engineering. Learn what prompts are, how AI models read them, and how to write clear instructions that get great results.
Prompt Engineering for Data Science & Analytics
Supercharge your data workflows with AI. 15 practical tutorials on using prompt engineering for data cleaning, EDA, machine learning, SQL, visualisation, and more.
Prompt Engineering for Image Generation
Turn words into stunning visuals. Master AI image generation tools like Midjourney, DALL·E 3, and Stable Diffusion with 18 focused tutorials — from first prompt to full brand identity.